People often report wanting to save more money than they actually do, and rigorous evidence has shown that simple reminders to save can be effective at helping people save more. Researchers working with IPA carried out evaluations in Ghana, Peru, the Philippines, and the Dominican Republic to build the evidence base about text-message reminders to save. In Ghana, the research team worked with Fidelity Bank to evaluate the impact of behaviorally targeted text messages on savings behavior. Preliminary results find that the reminders to save led people to use their bank accounts more frequently, but did not lead to increased savings, and that varying the timing and content of the messages did not impact the results.

Policy Issue 

People often report wanting to save more money than they actually do. There are a variety of potential reasons for this: misleading or incomplete information about savings products may prevent people from choosing the best one, they may not anticipate how changes in their financial situation could change their savings needs in the future, the human tendency to value instant gratification over future benefits may lead them to spend more in the short run, or they may simply forget to set aside money.1

Rigorous evidence has shown that simple reminders to save can be effective at helping people save more. For example, a study conducted with three financial institutions in Peru, Bolivia and the Philippines found that individuals randomly assigned to receive a monthly reminder via text message or letter deposited more money in savings accounts than a comparison group that did not receive reminders.2 Text message are a particularly promising way to deliver these reminders, with over 93 percent of adults in low- and middle-income countries having access to an active SIM card.3 However, more evidence is needed to better understand how text message reminders work in different contexts and better understand the mechanisms by which they may affect financial behavior.

This study in Ghana was carried out in tandem with other messaging projects in Peru, the Philippines, and the Dominican Republic.

Context of the Evaluation 

Fidelity Bank, the Ghanaian bank that served as a partner in this study, offers a financial product called a Smart Savings Account to its clients across the country. The account requires a form of national ID to open and has a minimum balance of GHS10 (about USD$1.70). The bank’s salespeople travel with account opening materials and a mobile phone to perform opening procedures, and independent retailers can both open accounts and perform agent banking services for clients.

Details of the Intervention 

Researchers working with IPA partnered with Fidelity Bank to evaluate the impact of behaviorally targeted text messages on account use and savings behavior. A total of 263,503 bank clients without an existing loan were assigned either to receive text messages (84,399 clients) or to a comparison group that did not receive the messages (179,104 clients).

The clients in the text messages group received one of six randomly-assigned variants:

  • Simple reminders: Received a straightforward reminder to save, such as “Don’t forget to save.”
  • Feature reminders: Received reminders about specific features of their savings accounts, such as the number of withdrawals available per month.
  • Transaction explanation reminders: Received reminders about how to deposit or withdraw money.
  • Other banking product reminders: Received reminders about other products offered by the bank, such as a microloan or micro-investment account.
  • Goal reminders: Received messages encouraging recipients to achieve their savings goals.
  • Social reminders: Received messages emphasizing the social acceptability of saving and encouraging recipients to join their peers who are saving.

In addition to content, the research team also randomized other elements of message delivery:

  • Time of day: Either early or late in the day.
  • Day of the week: A day between Monday and Saturday.
  • Frequency: Either once or twice per month.
  • Duration: Either six or nine months.
  • Personalization: Either a greeting with the client’s name at the beginning or a generic “Dear Client.”
  • Introduction: Either with or without an introduction.
  • Week within cycle: Either the first, second, third, or fourth week of the month.
  • Content: Some clients received the same message each month, while for others the type of message varied from month to month.

Researchers used the bank’s administrative data to assess the impacts of different forms of reminder delivery on account usage and savings.

Results and Policy Lessons 

Note: these results are preliminary and may change after further data analysis.

The text message reminders to save led people to use their bank accounts more frequently. However, while people who received messages made more transactions, they did not save more money, and they did not form account use habits that persisted after messages ended.

The increased use of bank accounts was especially pronounced for people who were not actively using their accounts before the messaging treatment started. Among these inactive users, 7.2 percent who received messages made a transaction, compared with 6.7 percent who did not.  indicating that messages may be more effective at encouraging people to start using their accounts rather than prompting them to use their accounts more.

Within the group that received reminders, most of the variations in how the reminders were delivered did not have significant impacts relative to a generic savings message. There were two exceptions for clients who had previously had an active account: reminders in the afternoon resulted in larger transactions than ones in the morning; and goal-based and persuasively-worded messages were less likely to encourage account activity.

Overall, the results suggest that reminder messages were effective at spurring inactive bank clients to use accounts more frequently, but that the content and timing of the messages was generally unimportant in affecting client behavior, and both withdrawals and deposits increased, suggesting that the increased activity did not result in increased savings.

Sources

1 For a discussion of recent empirical literature of these behavioral biases, see: Karlan, D., Ratan, A. L., & Zinman, J. (2014). Savings by and for the Poor: A Research Review and Agenda. Review of Income and Wealth60(1), 36-78.

2 Karlan, D., McConnell, M., Mullainathan, S., & Zinman, J. (2016). Getting to the top of mind: How reminders increase saving. Management Science.

3 ITU: Key ICT Statistics 2005-2016. Available at: http://www.itu.int/en/ITU-D/Statistics/Documents/statistics/2016/ITU_Key_2005-2016_ICT_data.xls (accessed 8/26/2016).